We describe a code designed for hot media {(T $\ge$} a few 10$^4$ K), optically thick to Compton scattering. It computes the structure of a plane-parallel slab of gas in thermal and ionization equilibrium, illuminated on one or on both sides by a given spectrum. Contrary to the other photoionization codes, it solves the transfer of the continuum and of the lines in a two stream approximation, without using the local escape probability formalism to approximate the line transfer. We stress the importance of taking into account the returning flux even for small column densities (10$^{22}$ cm$^{-2}$), and we show that the escape probability approximation can lead to strong errors in the thermal and ionization structure, as well as in the emitted spectrum, for a Thomson thickness larger than a few tenths. The transfer code is coupled with a Monte Carlo code which allows to take into account Compton and inverse Compton diffusions, and to compute the spectrum emitted up to MeV energies, in any geometry. Comparisons with {\sc{cloudy}} show that it gives similar results for small column densities. Several applications are mentioned.

Seyfert galaxies have been shown to exhibit a large variety of features in their X-ray spectra from which the environment of the central engine can be deduced. We focus on the two following aspects: the Warm Absorber, mainly responsible of the soft X-ray properties, and the reprocessing/reflecting plasma medium at the origin of the iron Kalpha fluorescent line. The physical parameters and the location of the so-called Warm Absorber (WA), a photoionized medium along the line of sight to the nuclear region, are more strongly constrained by optical coronal lines than by the oxygen edges observed in the soft X-rays and produced by the WA. The photoionization models also predict the intensities of the X-ray emission lines which are going to be detected with the new generation of X-ray satellites. An alternative model to the relativistic accretion disc is proposed to explain the profile of the X-ray iron Kalpha line observed in the Seyfert 1. This line can be formed in the framework of a quasi-spherical accretion of optically thick clouds. An optically thick photoionization code coupled with a Monte-Carlo code has been developed to compute the entire spectrum from the IR to the hard X-rays for a close geometry with a large covering factor. The multiple Compton reflections allow to reproduce the redshifted broad iron line as detected in several Seyfert 1 galaxies and in MCG-6-30-15.

The main characteristics of the average spectrum of radio quiet AGN in the UV and X-ray range are reviewed, and the emission mechanisms are discussed in the framework of accretion disk models, in particular the ``irradiated cold relativistic disk". It is shown that some problems arise in confronting the predictions of the model to the observations. We propose an alternative model in terms of a hot disk surrounded by a cold Compton thick medium. Finally we mention problems with remote regions of the accretion disk.

We propose an explanation of the variability of AGN based on a cloud model of accretion onto a black hole. These clouds, present at a distance of 10-100 Schwarzschild radii, possibly come from the disrupted innermost disk and they can partially obscure the central X-ray source due to their extreme Compton optical depth. We consider the implications of this scenario using a toy model of AGN spectra and support the results with full radiative transfer computations from the codes TITAN and NOAR. We show that small random rearrangements of the cloud distribution can happen on timescales of the order of 10^2 - 10^6 s and may lead to relatively high variability amplitude in the X-ray band if the mean covering factor is large. The normalized variability amplitude in the UV band is either the same as in the X-ray band, if the contribution from the dark sides of the clouds to the UV band is negligible, or it is smaller. The X-ray spectrum should basically preserve its spectral shape even during the high amplitude variability, within the frame of our model.

There are strong suggestions that the disk-like accretion flow onto massive black hole in AGN is disrupted in its innermost part (10-100 Rg), possibly due to the radiation pressure instability. It may form a hot optically thin quasi spherical (ADAF) flow surrounded by or containing denser clouds due to the disruption of the disk. Such clouds might be optically thick, with a Thompson depth of order of 10 or more. Within the frame of this cloud scenario (Collin-Souffrin et al. 1996, Czerny & Dumont 1998), obscuration events are expected and the effect would be seen as a variability. We consider expected random variability due to statistical dispersion in location of clouds along the line of sight for a constant covering factor. We discuss a simple analytical toy model which provides us with the estimates of the mean spectral properties and variability amplitude of AGN, and we support them with radiative transfer computations done with the use of TITAN code of Dumont, Abrassart & Collin (1999) and NOAR code of Abrassart (1999).